Target packet loss rate acquisition method and device, equipment and storage medium

Abstract

The invention discloses a target packet loss rate obtaining method and device, equipment and a storage medium, and relates to the technical field of communication, and the method comprises the steps: obtaining a sample packet loss rate set based on each candidate packet loss rate; determining a server stability index corresponding to the first sample packet loss rate according to the first sample packet loss rate; sequentially inputting server stability indexes corresponding to the sample packet loss rates in each sample packet loss rate set into a stability analysis model to obtain an evaluation factor combination; determining a target sample packet loss rate set according to the evaluation factor combination of the server stability index corresponding to each sample packet loss rate set; and taking the candidate packet loss rate corresponding to the target sample packet loss rate set as a target packet loss rate. The method can accurately quantify the local influence of the packet loss rate on the stability of the cloud product. Through quantification of the packet loss rate, data support is provided for optimal allocation of resources and network topology design, and the cloud service quality is improved. Risks are identified in advance, and system stability is guaranteed.

Description

Technical Field

[0001] This application relates to the field of network communication technologies, and particularly to a method, apparatus, device, and storage medium for obtaining a target packet loss rate. Background Art

[0002] Network packet loss is one of the key factors affecting network communication performance. Common reasons for network packet loss include network congestion, bandwidth limitation, hardware failures, software errors, etc. Network packet loss can cause transmission delays, throughput degradation, and affect service real-time performance. For cloud platforms, network packet loss leads to unstable server operation, lagging business operations of virtual machines and container instances, etc., greatly reducing the user experience, resulting in user loss and economic losses.

[0003] Currently, most cloud server packet loss detections are carried out through some network detection tools. These detection tools are mainly used for detecting network fault links and cannot quantify the impact of network packet loss rate on the stability of cloud products. Moreover, some current detection means are mainly carried out after a fault is discovered, lacking prediction tools. In addition, cloud products are complex and diverse, and there is currently no unified tool to evaluate the impact of network packet loss on server stability. Summary of the Invention

[0004] This application provides a method, apparatus, device, and storage medium for obtaining a target packet loss rate to at least solve a series of problems in the related art, such as that the cloud server packet loss detection means cannot quantify the impact of network packet loss rate on the stability of cloud products, mainly carried out after a fault is discovered and cannot be predicted, and there is currently no unified tool to evaluate the impact of network packet loss on server stability.

[0005] This application provides a method for obtaining a target packet loss rate, including:

[0006] Based on each candidate packet loss rate in a pre-determined set of candidate packet loss rates, obtain a set of sample packet loss rates within a preset packet loss rate range with each candidate packet loss rate as a benchmark;

[0007] According to the first sample packet loss rate in the first set of sample packet loss rates, determine the server stability index corresponding to the first sample packet loss rate, where the first set of sample packet loss rates is any one of multiple sets of sample packet loss rates, and the first sample packet loss rate is any sample packet loss rate in the first set of sample packet loss rates;

[0008] Input the server stability indexes corresponding to the sample packet loss rates in each set of sample packet loss rates into a pre-constructed stability analysis model in sequence, and obtain an evaluation factor combination of the server stability indexes corresponding to each set of sample packet loss rates;

[0009] Determine the target sample packet loss rate set according to the evaluation factor combinations of the server stability indicators respectively corresponding to each sample packet loss rate set.

[0010] Use the candidate packet loss rate corresponding to the target sample packet loss rate set as the target packet loss rate.

[0011] This application also provides a target packet loss rate acquisition device, including:

[0012] An acquisition module, configured to acquire a sample packet loss rate set within a preset packet loss rate range with each candidate packet loss rate as a benchmark, based on each candidate packet loss rate in a predetermined set of candidate packet loss rates.

[0013] A processing module, configured to determine a server stability indicator corresponding to the first sample packet loss rate according to the first sample packet loss rate in the first sample packet loss rate set, where the first sample packet loss rate set is any one of multiple sample packet loss rate sets, and the first sample packet loss rate is any sample packet loss rate in the first sample packet loss rate set.

[0014] An evaluation module, configured to sequentially input the server stability indicators respectively corresponding to the sample packet loss rates in each sample packet loss rate set into a pre-constructed stability analysis model, and acquire an evaluation factor combination of the server stability indicator corresponding to each sample packet loss rate set.

[0015] The processing module is further configured to determine a target sample packet loss rate set according to the evaluation factor combinations of the server stability indicators respectively corresponding to each sample packet loss rate set; and use the candidate packet loss rate corresponding to the target sample packet loss rate set as the target packet loss rate.

[0016] This application also provides a computer-readable storage medium, in which a computer program is stored, and when the computer program is executed by a processor, the steps of any of the above target packet loss rate acquisition methods are implemented.

[0017] This application also provides a computer program product, including a computer program, and when the computer program is executed by a processor, the steps of any of the above target packet loss rate acquisition methods are implemented.

[0018] Through this application, by analyzing each candidate packet loss rate individually, a deeper understanding of the server's performance at different packet loss rates can be achieved, thus enabling personalized analysis. By obtaining the set of sample packet loss rates within the preset packet loss rate range, the representativeness of the analysis samples can be ensured, avoiding the deviation of a single data point. Determining the server stability metrics corresponding to each sample packet loss rate helps identify the key factors affecting server stability. Inputting the stability metrics corresponding to the sample packet loss rates into the stability analysis model can optimize the model to more accurately predict and evaluate server stability. Determining the set of target sample packet loss rates through the evaluation factor combination helps identify the packet loss rate range most likely to affect server stability. Through analysis, the risks at different packet loss rates can be identified and evaluated, which helps formulate corresponding risk management measures. In addition, this method accurately quantifies the local impact of packet loss rate on the stability of cloud products by introducing a regression discontinuity design. Through the quantification of packet loss rate, data support is provided for the optimal allocation of resources and network topology design, improving the quality of cloud services. At the same time, data support is provided for cloud platform monitoring. When the packet loss rate reaches the target packet loss rate, risks can be identified in advance, alarms can be reported in real time and operations and maintenance personnel can be notified, or load balancing can be triggered to optimize network resources, ensuring the stability of the system and enhancing the user experience. This method has the advantages of high efficiency, low cost, and accurate quantification, and is applicable to various cloud products (such as cloud servers, cloud storage, cloud databases, etc.) and network environments, with broad application prospects. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] To more clearly illustrate the embodiments of the present application, the accompanying drawings required for use in the embodiments will be briefly introduced below. Obviously, the accompanying drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0020] Figure 1 Schematic flowchart of a method for obtaining a target packet loss rate provided by an embodiment of the present application;

[0021] Figure 2 Schematic flowchart of another method for obtaining a target packet loss rate provided by an embodiment of the present application;

[0022] Figure 3 Schematic flowchart of another method for obtaining a target packet loss rate provided by an embodiment of the present application;

[0023] Figure 4 Schematic flowchart of another method for obtaining a target packet loss rate provided by an embodiment of the present application;

[0024] Figure 5 Schematic flowchart of a method for performing sample sampling on each candidate breakpoint to obtain n sets of sample packet loss rates provided by an embodiment of the present application;

[0025] Figure 6 A schematic diagram of the data distribution after fitting provided by an embodiment of the present application;

[0026] Figure 7 A schematic diagram of the structure of a target packet loss rate acquisition device provided by an embodiment of the present application;

[0027] Figure 8 A schematic diagram of the structure of an electronic device provided by an embodiment of the present application. Detailed implementation manners

[0028] Next, the technical solutions in the embodiments of the present application will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the protection scope of the present application.

[0029] It should be noted that in the description of the present application, the terms "include", "comprise" or any other variant thereof are intended to cover a non-exclusive inclusion, so that a process, method, article or device including a series of elements not only includes those elements, but also includes other elements not expressly listed, or further includes elements inherent to such process, method, article or device. The terms "first", "second", etc. in the present application are used to distinguish similar objects, rather than to describe a specific order or sequence.

[0030] In order to enable those skilled in the art of the present technology to better understand the solution of the present application, the present application will be further described in detail below in conjunction with the accompanying drawings and specific implementation manners.

[0031] Network packet loss is one of the key factors affecting network communication performance, and it may occur in any transmission process between the source and the destination. Common reasons for network packet loss include network congestion, bandwidth limitation, hardware failure, software error, etc. Network packet loss will cause transmission delay, throughput decline, and affect service real-time performance. For cloud platforms, network packet loss causes unstable operation of servers, and lags in the operation of virtual machines and container instances, etc., greatly reducing the user experience, resulting in user loss and economic losses. Conducting a preliminary assessment of the impact of network packet loss on the platform in advance can prevent and solve potential network problems, optimize and allocate resources in advance, optimize the network layout, improve the data transmission speed, and thus improve the service quality and enhance user satisfaction.

[0032] Currently, packet loss detection in cloud servers mostly relies on some network detection tools. These detection tools are mainly used for detecting network fault links and cannot quantify the impact of network packet loss rate on the stability of cloud products. Moreover, some current detection methods are mainly carried out after a fault is discovered, lacking prediction tools. In addition, cloud products are complex and diverse, and there is currently no unified tool to evaluate the impact of network packet loss on server stability.

[0033] To solve the above problems, an embodiment of the present application provides a method for obtaining a target packet loss rate. Specifically, refer to Figure 1 As shown, the method includes the following steps:

[0034] Step S101: Based on each candidate packet loss rate in a pre-determined set of candidate packet loss rates, obtain a set of sample packet loss rates within a preset packet loss rate range with each candidate packet loss rate as a benchmark.

[0035] Specifically, since it is impossible to determine the target packet loss rate of network packet loss (the threshold of the impact of packet loss rate on the stability of cloud products such as cloud platforms, that is, the breakpoint value), a set of candidate breakpoint sets is defined according to the standards of network performance and data distribution here. The candidate breakpoint set is a set of estimated breakpoints. Then, for each breakpoint in the breakpoint set, obtain a set of sample packet loss rates within a preset packet loss rate range with each candidate packet loss rate as a benchmark.

[0036] Specifically, assume that the expression of the set of sample packet loss rates is as follows:

[0037] R = (r1, r2, r3, …, rn) T is a vector of n×1, where ri is the i-th sample packet loss rate in the set of sample packet loss rates. Among them, the larger n is, the more accurate the value of the finally obtained target packet loss rate is.

[0038] Step S102: Determine the server stability index corresponding to the first sample packet loss rate according to the first sample packet loss rate in the first set of sample packet loss rates.

[0039] In an optional example, the first sample packet loss rate can be input into the corresponding fault injection tool, and the fault injection tool is run to perform the corresponding packet loss operation according to the first sample packet loss rate, and then the server stability index corresponding to the first sample packet loss rate is obtained. Among them, the first set of sample packet loss rates is any set of sample packet loss rates among multiple sets of sample packet loss rates, and the first sample packet loss rate is any sample packet loss rate in the first set of sample packet loss rates.

[0040] Step S103: Input the server stability indexes corresponding to the sample packet loss rates in each set of sample packet loss rates into the pre-constructed stability analysis model in turn, and obtain the evaluation factor combination of the server stability index corresponding to each set of sample packet loss rates.

[0041] Specifically, assume that the set of sample packet loss rates corresponding to each candidate packet loss rate is Ra = (r1, r2, r3, …, rn). T . Input the server stability indicators corresponding to each packet loss rate ri in Ra into the stability analysis model. Based on the internal operations of the stability analysis model, an evaluation factor set of the server stability indicators corresponding to the set of sample packet loss rates can be obtained.

[0042] In a specific example, the stability analysis model can be, for example, a linear regression analysis model.

[0043] Optionally, the expression of the stability analysis model is as follows, for example:

[0044] Y = Mβ + ∈ (Formula 1)

[0045] Among them, Y represents the server stability indicator, and the matrix M = (C, R T , D, N);

[0046] C represents the intercept, and R T represents the packet loss rate corresponding to the server stability indicator, D represents the processing variable, and when the packet loss rate of the i-th sample ≥ the candidate packet loss rate value, then Di = 1, or when the packet loss rate of the i-th sample < the candidate packet loss rate value, then Di = 0, N represents the interaction term, and the interaction term Ni corresponding to the packet loss rate of the i-th sample = the packet loss rate of the i-th sample × the processing variable, β represents the evaluation factor combination, and ∈ represents the error vector.

[0047] In this formula, one server indicator can be input each time to obtain an expression of the stability analysis model, and then finally, based on the expressions of multiple stability analysis models, all the evaluation factors in the evaluation factor combination can be jointly obtained. Therefore, regardless of whether ∈ is an unknown vector, as long as the number of collected sample packet loss rates is set sufficiently, it can be calculated. For example, if the number of collected sample packet loss rates is set to an even number, then ∈ can be directly eliminated. Or, even if it is set to an odd number, as long as the number of collected sample packet loss rates is set to be greater than or equal to the total number of evaluation factors in the evaluation factor combination and the error vector, it is okay.

[0048] Step S104, determine the target sample packet loss rate set according to the evaluation factor combination of the server stability indicators corresponding to each set of sample packet loss rates.

[0049] Specifically, the evaluation factor combination itself is used to indicate the change trend of the server stability indicator as the packet loss rate changes, and can also be understood as the slope of the change of the server stability indicator as the packet loss rate increases. The larger the value of the slope, the greater the change trend.

[0050] In an alternative example, the evaluation factor combination includes a first evaluation factor used to indicate the server stability indicators corresponding to the sample packet loss rates less than the candidate packet loss rate in the sample packet loss rate set respectively, a second evaluation factor used to indicate the server stability indicators corresponding to the sample packet loss rates greater than the candidate packet loss rate in the sample packet loss rate set respectively, and a third evaluation factor corresponding to the candidate packet loss rate.

[0051] In an alternative example, it can be implemented through the following method steps. For details, please refer to Figure 2 As shown in the figure, it includes:

[0052] Step S201: Input each group of evaluation factor combinations into the pre-constructed performance prediction model in sequence to obtain the target evaluation factor combination.

[0053] Step S202: Determine the target sample packet loss rate set according to the target evaluation factor combination.

[0054] Specifically, the first to third evaluation factors can be jointly input into the python algorithm module. For example, the R 2 (R-squared, coefficient of determination) algorithm in the python algorithm is used to obtain the evaluation factor combination corresponding to the model with the largest R2 as the target evaluation factor combination.

[0055] And determine the sample packet loss rate set corresponding to the target evaluation factor combination as the target sample packet loss rate set.

[0056] Step S105: Use the candidate packet loss rate corresponding to the target sample packet loss rate set as the target packet loss rate.

[0057] Specifically, since the target sample packet loss rate set is determined based on the target candidate packet loss rate, the target candidate packet loss rate is the target packet loss rate.

[0058] A method for obtaining the target packet loss rate provided by the embodiment of the present application can perform individual analysis on each candidate packet loss rate, enabling a deeper understanding of the server's performance at different packet loss rates, thereby achieving personalized analysis. By obtaining the set of sample packet loss rates within the preset packet loss rate range, the representativeness of the analysis samples can be ensured, avoiding the deviation of a single data point. Determining the server stability indicators corresponding to each sample packet loss rate helps identify the key factors affecting server stability. Inputting the stability indicators corresponding to the sample packet loss rates into the stability analysis model can optimize the model to more accurately predict and evaluate server stability. Determining the set of target sample packet loss rates through the evaluation factor combination helps identify the packet loss rate range most likely to affect server stability. Through analysis, the risks at different packet loss rates can be identified and evaluated, which helps formulate corresponding risk management measures. In addition, this method accurately quantifies the local impact of the packet loss rate on the stability of cloud products by introducing a regression discontinuity design. Through the quantification of the packet loss rate, it provides data support for the optimal allocation of resources and network topology design, improving the quality of cloud services. At the same time, it provides data support for cloud platform monitoring. When the packet loss rate reaches the target packet loss rate, it can identify risks in advance, report alarms in real time and notify the operation and maintenance personnel, or trigger load balancing to optimize network resources, ensuring the stability of the system and enhancing the user experience. This method has the advantages of high efficiency, low cost, and accurate quantification, and is applicable to various cloud products (such as cloud servers, cloud storage, cloud databases, etc.) and network environments, having a wide range of application prospects.

[0059] In an alternative example, based on each candidate packet loss rate in a pre-determined set of candidate packet loss rates, a set of sample packet loss rates within the preset packet loss rate range with each candidate packet loss rate as the benchmark is obtained. The specific method steps are as follows:

[0060] Taking each candidate packet loss rate as the center respectively, determine the packet loss rates within the preset range on both the left and right sides of the center as the set of sample packet loss rates within the preset packet loss rate range with each candidate packet loss rate as the benchmark.

[0061] In a specific example, for example, for each pre-determined candidate packet loss rate t, select n packet loss rates within the range [t - 2, t + 2] as the n sample packet loss rates in the set of sample packet loss rates corresponding to this breakpoint. Therefore, when there are n candidate packet loss rates, the corresponding set of sample packet loss rates is also n.

[0062] Based on the foregoing embodiment, according to the first sample packet loss rate in the first set of sample packet loss rates, determine the server stability indicator corresponding to the first sample packet loss rate. The specific method steps are as follows. For details, please refer to Figure 3 as shown:

[0063] Step S301, automatically trigger a fault injection instruction according to the first sample packet loss rate.

[0064] Step S302: Obtain the logs during the execution phase of the fault injection instruction, analyze and process the logs, and obtain the server stability index corresponding to the first sample packet loss rate.

[0065] Among them, the fault injection instruction is used to indicate to perform a packet loss operation according to the first sample packet loss rate and the packet loss time corresponding to the first sample packet loss rate.

[0066] Specifically, input the first sample packet loss rate ri of the packet loss rate, and automatically trigger fault injection. Then run the fault injection tool to discard a certain proportion of data packets on the corresponding network card according to the preset packet loss rate ri and the pre-configured packet loss time. After the fault injection, start the fault detection tool, and this detection tool sends multiple Ping packets to confirm whether the fault is actually injected by verifying the packet loss field of the packets.

[0067] Within the packet loss time range of the fault tool, start the server metric collection script. This script will continuously run (Application Programming Interface, API for short) during the fault, collect the logs of the API before the end of the fault, parse and refine the log content, convert the data into a standard format, calculate the request success rate of all APIs during the fault, and assign it to the server stability index yi corresponding to the packet loss rate ri.

[0068] Optionally, based on the foregoing method steps, the method may further include:

[0069] After the server stability index is collected, stop the fault injection script, and call the fault detection tool again to send multiple Ping packets to verify whether the fault is successfully destroyed by checking the packet loss field of the packets.

[0070] In the above method, by simulating the packet loss situation in a real network environment, the stability and fault recovery ability of the server can be evaluated more realistically. Automatically triggering the fault injection instruction can automate the testing process, improve testing efficiency and consistency. By analyzing the log data during the fault injection process, quantitative server stability indicators can be obtained to provide data support for decision-making. More specifically, by precisely controlling the packet loss rate and packet loss time, the server performance under specific network conditions can be simulated more accurately. By analyzing the logs during the fault injection process, server stability problems can be quickly located, which helps to quickly fix the problems. According to the collected stability indicators, system bottlenecks can be identified, and thus optimization and adjustment can be carried out. Through the above method, the performance of the server under different packet loss rates can be analyzed to evaluate the risk tolerance of the system. Through fault injection testing, the robustness and fault tolerance of the system in the face of network instability can be enhanced. By continuously evaluating and optimizing server stability, the quality of service and user experience can be ultimately improved.

[0071] Based on any of the foregoing embodiments, the method may further include the following method steps. For details, see Figure 4 As shown, it includes:

[0072] Step S401, periodically obtain the topology information of the environment where the stability analysis model is located.

[0073] Step S402, when it is determined that the topology information has changed, generate a trigger mechanism.

[0074] Step S403, according to the trigger mechanism, re-determine the target packet loss rate.

[0075] Specifically, due to the complex and changeable cloud platform environment, affected by various conditions such as the hardware environment and implementation scheme, each environment has its uniqueness, and the topology of the same environment will also change continuously. In the present invention, the environment topology information of the cloud platform will be automatically parsed, the topology environment changes will be monitored in real time, after perceiving the topology environment change, the data collection process will be automatically triggered, the server stability indicators will be updated, and the fitting breakpoint regression model will be reconstructed to obtain the optimal target breakpoint. At the same time, it will be checked whether the packet loss rate threshold setting of the monitoring module network card is greater than the current target breakpoint. If it is greater than the target breakpoint, the packet loss rate threshold setting will be updated to prevent losses caused by untimely detection.

[0076] Specifically, it can be understood that the change of the topology information is an opportunity to trigger the change of the network packet loss rate threshold. Therefore, once the topology information change is detected, the operation of obtaining the network packet loss rate threshold will be executed again.

[0077] In an alternative embodiment, the solution for obtaining the topology information can be as follows:

[0078] 1. Data collection: Collect network traffic data, including information such as IP addresses, port numbers, and packet sizes, through sensors or agents deployed in the network.

[0079] 2. Feature extraction: Preprocess the collected data and extract key features reflecting the network topology structure, such as traffic patterns and node activity levels.

[0080] 3. Model training: Use machine learning algorithms, such as deep learning or support vector machines, to learn the extracted features and establish a network topology model.

[0081] 4. Topology recognition: Apply the trained model to actual network traffic data to identify key nodes and connections in the network, thereby obtaining network topology information.

[0082] In addition, to improve the accuracy and reliability of network topology information, we can introduce the following mechanisms:

[0083] 1. Multi-source data fusion: Integrate network traffic data from different sources to improve the comprehensiveness and accuracy of topology information.

[0084] Specifically, collect data through multiple data sources and perform steps such as data cleaning, preprocessing, and feature extraction to achieve standardized and consistent processing of the data, thereby improving the integrity and precision of network topology information.

[0085] 2. Real-time monitoring and warning: Monitor changes in network traffic in real time, issue warnings for abnormal situations, and adjust the network topology model in a timely manner.

[0086] Specifically, use traffic monitoring tools (such as Prometheus, Nagios, Zabbix, etc.) to collect network traffic data. Monitor key metrics, such as inbound and outbound traffic, packet size, error rate, packet loss rate, etc. Detect abnormal traffic patterns. For example, a sudden large traffic peak may be a sign of a DDoS attack. In addition, an increase in the network device error rate may be a signal of hardware failure or misconfiguration. Unusually large data packets may also be a sign of malware activity or a misconfigured load balancer. These situations all indicate possible abnormal traffic patterns.

[0087] When the monitoring tool detects an anomaly, it triggers an alert (such as an email, SMS, Slack message, etc.) to notify the administrator. After receiving the alert, the administrator first analyzes the detailed information of the alert, including the time, involved IP addresses, ports, etc. Restrict the abnormal traffic. For example, use firewall rules to filter specific IP addresses or traffic types. Among them, if the traffic is too large and causes a bottleneck, it is possible to redirect the traffic to other network paths or use a load balancer to disperse the traffic. Or, if a certain device is suspected to have problems, it can be isolated from the network to prevent further impact. Finally, according to the analysis results, the following adjustments may be required to the network topology:

[0088] 1) Upgrade the bandwidth for the bottleneck link;

[0089] 2) Change the routing path: If there is a problem with a certain AS path, the traffic can be re-routed to avoid this path.

[0090] 3) If it is determined that there is a hardware failure, the faulty device needs to be replaced.

[0091] 4) If the anomaly is caused by a configuration error, the configuration of the network device needs to be corrected.

[0092] Finally, after adjusting the network topology, continue to monitor the traffic and performance metrics to verify the effect of the adjustment.

[0093] 3. Node reputation assessment: Conduct a reputation assessment on the nodes in the network, screen out the trustworthy nodes, and improve the reliability of the topology information.

[0094] The specific implementation process includes, for example:

[0095] 1) Collect data from each node;

[0096] 2) Use a pre-built scoring model to score the nodes;

[0097] 3) Screen out the trustworthy nodes according to the pre-determined scoring criteria;

[0098] 4) Update the network topology and place the trustworthy nodes in key positions.

[0099] 5) Continuously monitor the node performance and collect new data.

[0100] 6) Adjust the scoring model and topology according to the new data.

[0101] Obtain the topology information in the above manner, and determine whether the topology information has changed. And generate a trigger mechanism when it is determined that the topology information has changed.

[0102] In the above method, by periodically obtaining topology information, it is possible to ensure that the data basis of the stability analysis model is up-to-date, thereby improving the accuracy of the analysis results. When the network topology changes, adjusting the target packet loss rate in a timely manner can ensure that the stability evaluation of the server or network system is more reliable. Moreover, the system can also quickly respond and adjust the analysis parameters, which is crucial for quickly discovering and solving problems. Adjusting the target packet loss rate according to the change of the network topology helps to optimize the allocation and utilization of network resources. By monitoring the topology changes, potential problems that may affect the system stability can be discovered in advance and measures can be taken to prevent them. By ensuring continuous monitoring of network stability and service quality, the user experience can be ultimately improved. Furthermore, by adapting to the changes in the network topology, the system's resilience and anti-interference ability can be enhanced.

[0103] Next, a specific embodiment will be used to illustrate the overall implementation process of the foregoing method steps. Specifically, refer to the following:

[0104] 1. Data collection.

[0105] Within the preset packet loss rate range selected based on the candidate packet loss rate, for example, within the range of [t - 2, t + 2] introduced above, collect the stability data of cloud products at different packet loss rates. Among them, the more samples are collected, the more accurate the data. In the embodiment of the present application, n groups of samples will be collected within each preset packet loss rate range.

[0106] The network packet loss of the cloud platform mainly affects the request success rate of platform access. Define the dependent variable Y = (y1, y2, y3,..., yn) T , where yi represents the server stability index corresponding to the i-th group of samples, that is, the platform access request success rate corresponding to ri packet loss rates.

[0107] Among them, the data collection process is as follows:

[0108] 1. For each packet loss rate sample corresponding to the candidate breakpoint, R = (r1, r2, r3,..., rn) T。Input the packet loss rate ri in the manner described above in sequence, and automatically trigger fault injection. The fault injection tool discards a certain proportion of data packets on the corresponding network card according to the preset packet loss rate and packet loss time. After the fault injection, start the fault detection tool. This detection tool sends multiple Ping messages and verifies whether the fault is actually injected by checking the packet loss field of the messages. Within the packet loss time range of the fault tool, start the server metric collection script. This script continuously runs the API during the fault, collects the API logs before the end of the fault, parses and refines the log content, converts the data into a standard format, calculates the request success rate of all APIs during the fault, and assigns it to the server stability metric yi corresponding to the packet loss rate ri. After the server stability metric collection is completed, stop the fault injection script, and call the fault detection tool again to send multiple Ping messages to verify whether the fault is successfully destroyed by checking the packet loss field of the messages.

[0109] 2. Perform the above process n times for each candidate breakpoint in the manner of step 1 to obtain the server stability metrics Ya=(y1, y2, y3,..., yn) corresponding to each set of sample packet loss rates T 。

[0110] The specific execution process is as follows Figure 5 shown, including:

[0111] 1. Initialize i = 1 and y = 0;

[0112] 2. If i ≤ n, input the packet loss rate ri;

[0113] 3. Trigger fault income;

[0114] 4. Verify the fault;

[0115] 5. Obtain and collect the server stability metrics;

[0116] 6. After incrementing i by 1, enter the next loop and continue to obtain the server stability analysis metrics;

[0117] 7. End the above operations when i is greater than n;

[0118] 8. Destroy the fault;

[0119] 9. End.

[0120] Perform the above loop operation once for each set of sample packet loss rates.

[0121] 3. Obtain the evaluation factor combination through the linear regression model.

[0122] Specifically, the linear regression model has been introduced in detail above, that is, Y = Mβ + ∈.

[0123] In this model, since Y(Ya) has been obtained in the previous text and is a known number, and M can be obtained through the data introduced in the previous text. Therefore, β can be solved by the least squares method as β=(M T M) -1 M T Y is a 4×1 vector group.

[0124] Specifically, in the 4×1 vector corresponding to β, β0 is the stability when the packet loss rate is 0, β1 is the trend change of stability with the increase of the packet loss rate in the control group (ri <= candidate packet loss rate), which is the first evaluation factor introduced in the previous text. β2 is the percentage of stability decrease after exceeding the breakpoint, which can also be understood as the third evaluation factor corresponding to the candidate packet loss rate. β3 is the trend change of stability with the increase of the packet loss rate after exceeding the breakpoint, which is the second evaluation factor introduced in the previous text. In addition, the evaluation factor combination can also include the stability indication parameter when the packet loss rate is 0, that is, the slope β0.

[0125] Taking a set of sample packet loss rate sets and the corresponding server stability index combinations as an example, the schematic diagram of the fitted data distribution is specifically as Figure 6 shown. Figure 6 is just an example diagram, and the specific actual diagram is determined according to the actual data. Among them, Figure 6 the abscissa in is the sample packet loss rate in the sample packet loss rate set, and the ordinate is the server stability index corresponding to each sample packet loss rate. Figure 6 The candidate packet loss rate set in is set to 5%.

[0126] From Figure 6 it can be seen that on the left side of the breakpoint, that is, in the range of [t - 2, t], as the sample packet loss rate increases, the server stability index also gradually decreases. However, in the range of [t, t + 2], since the packet loss rate has exceeded the candidate packet loss rate, the trend of the server stability index decreasing is more obvious. In addition, Figure 6 also shows β1~β3, β0 ( Figure 6 not shown in) is infinitely close to Figure 6 the leftmost point on the left diagonal line in.

[0127] 4. Determine the target sample packet loss rate set.

[0128] Specifically, first determine the target evaluation factor combination according to the foregoing method steps. For example, according to each group of evaluation factor combinations, calculate R 2 (which can be implemented through a python module), and select the evaluation factor combination with the largest R 2 as the target evaluation factor combination; then, select the sample packet loss rate combination that generates the server stability index corresponding to the target evaluation factor combination as the target sample packet loss rate combination.

[0129] Further optionally, it further includes: 5. Update the target packet loss rate.

[0130] Specifically, as introduced above, due to the complex and ever-changing cloud platform environment, affected by various conditions such as the hardware environment and implementation scheme, each set of environments has its uniqueness, and the same set of environment topologies will also change continuously. Therefore, in this application, it also includes generating a trigger mechanism according to the environment topology information to update the target packet loss rate. The specific implementation process has been described in detail above, so it will not be elaborated here too much. Through the description of the above embodiments, those skilled in the art can clearly understand that the method according to the above embodiments can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware, but in many cases, the former is a better implementation method.

[0131] The embodiment of this application also provides a target packet loss rate acquisition device, specifically referring to Figure 7 as shown, the device includes: an acquisition module 701, a processing module 702, and an evaluation module 703.

[0132] The acquisition module 701 is used to obtain a set of sample packet loss rates within a preset packet loss rate range based on each candidate packet loss rate in a pre-determined set of candidate packet loss rates; the processing module 702 is used to determine a server stability index corresponding to the first sample packet loss rate according to the first sample packet loss rate in the first sample packet loss rate set, where the first sample packet loss rate set is any one of multiple sample packet loss rate sets, and the first sample packet loss rate is any sample packet loss rate in the first sample packet loss rate set;

[0133] The evaluation module 703 is used to sequentially input the server stability indexes corresponding to the sample packet loss rates in each sample packet loss rate set into a pre-constructed stability analysis model to obtain an evaluation factor combination of the server stability indexes corresponding to each sample packet loss rate set;

[0134] The processing module 702 is further used to determine a set of target sample packet loss rates according to the evaluation factor combination of the server stability indexes corresponding to each sample packet loss rate set; and use the candidate packet loss rate corresponding to the set of target sample packet loss rates as the target packet loss rate.

[0135] In an optional example, the acquisition module 701 is specifically used to determine the packet loss rates within a preset range on the left and right of the center respectively with each candidate packet loss rate as the center, as the set of sample packet loss rates within the preset packet loss rate range based on each candidate packet loss rate.

[0136] In an optional example, the processing module 702 is specifically configured to automatically trigger a fault injection instruction according to the first sample packet loss rate, where the fault injection instruction is used to indicate performing a packet loss operation according to the first sample packet loss rate and the packet loss time corresponding to the first sample packet loss rate;

[0137] Obtain the log during the execution stage of the fault injection instruction, analyze and process the log, and obtain the server stability index corresponding to the first sample packet loss rate.

[0138] In an optional example, the evaluation module 703 is specifically configured to sequentially input each group of evaluation factor combinations into a pre-constructed performance prediction model to obtain a target evaluation factor combination;

[0139] Determine a target sample packet loss rate set according to the target evaluation factor combination.

[0140] In an optional example, the evaluation factor combination includes a first evaluation factor used to indicate the server stability index determined according to the sample packet loss rates less than the candidate packet loss rate in the sample packet loss rate set, a second evaluation factor used to indicate the server stability index corresponding to the sample packet loss rates greater than the candidate packet loss rate in the sample packet loss rate set, and a third evaluation factor corresponding to the candidate packet loss rate.

[0141] In an optional example, the processing module 702 is further configured to periodically obtain the topology information of the environment where the stability analysis model is located;

[0142] When it is determined that the topology information has changed, generate a trigger mechanism;

[0143] Re-determine the target packet loss rate according to the trigger mechanism.

[0144] In an optional example, the expression of the stability analysis model is as follows:

[0145] Y = Mβ + ∈ (Formula 2)

[0146] Wherein, Y represents the server stability index, the matrix M = (C, R T , D, N);

[0147] C represents the intercept, R T represents the packet loss rate corresponding to the server stability index, D represents the processing variable, and when the i-th sample packet loss rate ≥ the candidate packet loss rate value, then Di = 1, or when the i-th sample packet loss rate < the candidate packet loss rate value, then Di = 0, N represents the interaction term, and the interaction term Ni corresponding to the i-th sample packet loss rate = the i-th sample packet loss rate × the processing variable, β represents the evaluation factor combination, and ∈ represents the error vector.

[0148] For the descriptions of the features in the corresponding embodiments of a target packet loss rate acquisition device provided by the embodiments of the present application, reference can be made to the relevant descriptions in the corresponding embodiments of the target packet loss rate acquisition method, which will not be elaborated here one by one.

[0149] A target packet loss rate acquisition device provided by the embodiments of the present application can perform separate analysis on each candidate packet loss rate, which can help to more deeply understand the performance of the server at different packet loss rates, thereby realizing personalized analysis. By obtaining a set of sample packet loss rates within a preset packet loss rate range, the representativeness of the analysis samples can be ensured, and the deviation of a single data point can be avoided. Determining the server stability indicators corresponding to each sample packet loss rate helps to identify the key factors affecting the server stability. Inputting the stability indicators corresponding to the sample packet loss rates into the stability analysis model can optimize the model to more accurately predict and evaluate the server stability. Determining the set of target sample packet loss rates through the evaluation factor combination helps to identify the packet loss rate range that is most likely to affect the server stability. Through analysis, the risks at different packet loss rates can be identified and evaluated, which helps to formulate corresponding risk management measures. In addition, this method accurately quantifies the local impact of the packet loss rate on the stability of cloud products by introducing a regression discontinuity design. Through the quantification of the packet loss rate, it provides data support for the optimal allocation of resources and the network topology design, improving the quality of cloud services. At the same time, it provides data support for cloud platform monitoring. When the packet loss rate reaches the target packet loss rate, it can identify risks in advance, report alarms in real time and notify the operation and maintenance personnel, or trigger load balancing to optimize network resources, ensuring the stability of the system and improving the user experience. This method has the advantages of high efficiency, low cost, accurate quantification, etc., and is applicable to various cloud products (such as cloud servers, cloud storage, cloud databases, etc.) and network environments, and has a wide application prospect.

[0150] The embodiments of the present application also provide an electronic device, such as Figure 8 shown, including a memory 10 and a processor 20. A computer program is stored in the memory 10, and the processor 20 is configured to run the computer program to execute the steps in any of the above embodiments of the target packet loss rate acquisition method.

[0151] The embodiments of the present application also provide a computer-readable storage medium, in which a computer program is stored. The computer program is configured to execute the steps in any of the above embodiments of the target packet loss rate acquisition method or execute the steps in any of the above embodiments of the data reading method when running.

[0152] In an exemplary embodiment, the above computer-readable storage medium may include, but is not limited to: various media that can store computer programs such as USB flash drives, read-only memory (ROM for short), random access memory (RAM for short), mobile hard disks, magnetic disks, or optical discs.

[0153] The embodiments of the present application also provide a computer program product. The above computer program product includes a computer program, and when the computer program is executed by a processor, the steps in any of the above embodiments of the target packet loss rate acquisition method are implemented.

[0154] The embodiments of the present application also provide another computer program product, including a non-volatile computer-readable storage medium. The non-volatile computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps in any of the above embodiments of the data reading method are implemented.

[0155] Those skilled in the art can further realize that the units and algorithm steps of each example described in combination with the embodiments disclosed herein can be implemented by electronic hardware, computer software, or a combination of the two. To clearly illustrate the interchangeability of hardware and software, the components and steps of each example have been generally described according to functions in the above description. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution. Skilled professionals can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of the present application.

[0156] The above has introduced in detail a target packet loss rate acquisition method, device, equipment, and storage medium provided by the present application. Specific examples are used in this article to elaborate on the principle and implementation manner of the present application. The description of the above embodiments is only used to help understand the method and its core idea of the present application. It should be noted that for those of ordinary skill in the art in the technical field, without departing from the principle of the present application, several improvements and modifications can be made to the present application, and these improvements and modifications also fall within the protection scope of the claims of the present application.

Claims

1. A method for obtaining an objective packet loss rate, characterized in that, The method includes: Based on each candidate packet loss rate in a predetermined set of candidate packet loss rates, obtain a set of sample packet loss rates within a preset packet loss rate range with each candidate packet loss rate as a benchmark; According to the first sample packet loss rate in the first set of sample packet loss rates, determine a server stability index corresponding to the first sample packet loss rate, where the first set of sample packet loss rates is any one of multiple sets of sample packet loss rates, and the first sample packet loss rate is any sample packet loss rate in the first set of sample packet loss rates; Input the server stability indexes corresponding to the sample packet loss rates in each set of sample packet loss rates into a pre-constructed stability analysis model in sequence, and obtain an evaluation factor combination of the server stability indexes corresponding to each set of sample packet loss rates; Determine a target set of sample packet loss rates according to the evaluation factor combination of the server stability indexes corresponding to each set of sample packet loss rates; Use the candidate packet loss rate corresponding to the target set of sample packet loss rates as the target packet loss rate.

2. The method according to claim 1, characterized in that, The step of, based on each candidate packet loss rate in a predetermined set of candidate packet loss rates, obtaining a set of sample packet loss rates within a preset packet loss rate range with each candidate packet loss rate as a benchmark specifically includes: Taking each candidate packet loss rate as a center respectively, determine the packet loss rates within a preset range on both the left and right of the center as the set of sample packet loss rates within a preset packet loss rate range with each candidate packet loss rate as a benchmark.

3. The method according to claim 1, wherein The step of, according to the first sample packet loss rate in the first set of sample packet loss rates, determining a server stability index corresponding to the first sample packet loss rate specifically includes: Automatically trigger a fault injection instruction according to the first sample packet loss rate, where the fault injection instruction is used to indicate to perform a packet loss operation according to the first sample packet loss rate and the packet loss time corresponding to the first sample packet loss rate; Obtain the logs during the execution stage of the fault injection instruction, and analyze and process the logs to obtain the server stability index corresponding to the first sample packet loss rate.

4. The method according to claim 3, wherein The step of, according to the evaluation factor combination of the server stability indexes corresponding to each set of sample packet loss rates, determining a target set of sample packet loss rates specifically includes: Input each group of the evaluation factor combinations into a pre-constructed performance prediction model in sequence to obtain a target evaluation factor combination; Determine the target set of sample packet loss rates according to the target evaluation factor combination.

5. The method according to claim 4, wherein The evaluation factor combination includes a first evaluation factor used to indicate the server stability indexes determined according to the sample packet loss rates less than the candidate packet loss rate in the set of sample packet loss rates, a second evaluation factor used to indicate the server stability indexes determined according to the sample packet loss rates greater than the candidate packet loss rate in the set of sample packet loss rates, and a third evaluation factor corresponding to the candidate packet loss rate.

6. The method according to any one of claims 1-5, characterized in that, The method further includes: Periodically obtain the topology information of the environment where the stability analysis model is located; When it is determined that the topology information has changed, generate a trigger mechanism; According to the trigger mechanism, re-determine the target packet loss rate.

7. The method according to claim 5, characterized in that The method according to any one of 1-5, characterized in that the expression of the stability analysis model is as follows: Y = Mβ + ∈ Among them, Y represents the server stability index, and the matrix M = (C, R T , D, N); C represents the intercept, and R T represents the packet loss rate corresponding to the server stability index, D represents the treatment variable, and when the packet loss rate of the i-th sample ≥ the candidate packet loss rate value, then Di = 1, or when the packet loss rate of the i-th sample < the candidate packet loss rate value, then Di = 0, N represents the interaction term, and the interaction term Ni corresponding to the packet loss rate of the i-th sample = the packet loss rate of the i-th sample × the treatment variable, β represents the evaluation factor combination, and ∈ represents the error vector.

8. An apparatus for obtaining a target packet loss rate, characterized in that The device includes: An acquisition module, configured to acquire a set of sample packet loss rates within a preset packet loss rate range based on each candidate packet loss rate in a pre-determined set of candidate packet loss rates; A processing module, configured to determine a server stability index corresponding to the first sample packet loss rate according to the first sample packet loss rate in the first set of sample packet loss rates, where the first set of sample packet loss rates is any one of multiple sets of sample packet loss rates, and the first sample packet loss rate is any sample packet loss rate in the first set of sample packet loss rates; An evaluation module, configured to sequentially input the server stability indexes corresponding to the sample packet loss rates in each set of sample packet loss rates into a pre-constructed stability analysis model, and acquire an evaluation factor combination of the server stability indexes corresponding to each set of sample packet loss rates; The processing module is further configured to determine a target set of sample packet loss rates according to the evaluation factor combination of the server stability indexes corresponding to each set of sample packet loss rates; and use the candidate packet loss rate corresponding to the target set of sample packet loss rates as the target packet loss rate.

9. An electronic device, characterized in that, Includes: A memory, configured to store a computer program; A processor, configured to implement the steps of the target packet loss rate acquisition method according to any one of claims 1 to 7 when executing the computer program.

10. A computer-readable storage medium, characterized in that, A computer program is stored in the computer-readable storage medium, where the computer program implements the steps of the target packet loss rate acquisition method according to any one of claims 1 to 7 when executed by a processor.

Images